Dell EMC Unity provides some significant improvements over the previous CLARiiON designs, including integrated block and file support, faster processors, and all flash. However, as with any platform it still has certain practical limitations.
In this short blog we will examine how implementation choices can affect the system’s overall capabilities.
The Dell EMC Unity is a derivation of the traditional CLARiiON based platform with beefier hardware. The system consists of the following components:
- Fibre Channel Ports (Up to 16 Gbps)
- Intel E5 Processors (Varies with model)
- 150-1500 Drives (Varies with model)
- 48-768 GB of memory (Varies with model)
- Storage Provisioning Pool with Volumes (VVol Support)
Measurement can be performed using the Unisphere UI/CLI or by collecting data from the SMI-S interface. The UI/CLI includes metrics from the following components:
- *File System
- *VVol Datastore
- *Fibre Channel Port
- *iSCSI interface
- *Ethernet Port
*These are not currently provided directly by the SMI-S interface
Viewing Health of the Dell EMC Unity System
Ideally you want to see the key areas that contribute to the performance of the system at a single glance. IntelliMagic Vision provides a simple way to view the health of your Dell EMC Unity systems in a storage system level dashboard as shown in Figure 1.
This dashboard shows the overall I/O rate, the read and write throughput, response time and drive busy for the entire system.
At the top level of Figure 1, we see that the read and write throughput and response times are healthy as indicated by the green square icons. The blue I/O rate bubble is intended to provide a view of the relative I/O rate for the storage system. The yellow warning sign over the Drive busy indicates that the disk drive utilization is higher than expected.
Drilling down from the system dashboard to the charts over time provides additional detail and clarity around the rationale behind the health of the system including the warning on the drive utilization.
The drill down to the time charts shows the drive utilization is fairly high as displayed in Figure 2.
The drive busy shows a rating of 0.47.
The rating that IntelliMagic Vision provides for rated charts indicates the level of risk for the associated device and metric as it relates to performance. During the peak period the estimated drive utilization was around 70%. There was some increase in the response time on the front-end but it was not significant at the current level of activity.
Nevertheless, it is good to understand that every piece of hardware has physical limitations on the number of requests it can service. In this case the devices are becoming saturated when the system achieves around 250,000 IOPS on the backend disk drives.
In addition to identifying any system bottlenecks, it is also important to be able to identify the host(s) driving the workloads. Within IntelliMagic Vision it is a matter of clicking on the Host Throughput chart as shown in Figure 3: Host Workloads.
You can see in the chart that the host(s) named lasvmc1s13__lasvmc1s2__lasvmc1s3__lasvmc2s13 has the workload that correlates with the I/O rate, throughput and drive utilization. This host is responsible for the spikes in the workload that occur throughout this analysis period.
Identifying Front-End Port Imbalances
In order to optimize the throughput capabilities of the Dell EMC Unity, it is extremely important to balance the throughput across the front-end storage ports. Figure 4 demonstrates the balance in front-end throughput to the active storage ports.
The balance chart in Figure 4 shows the average as represented by the green dot, the 10th and 90th percentiles as represented by the green rectangle and the minimum and maximum values as represented by the yellow rectangle.
Ports spa_fc4 and spa_fc5 are operating at a much higher level than ports spb_fc4 and spb_fc5. Ideally you want to balance the workload across the two controllers (SPA & SPB).
For VMware ESXi systems the preferred host multi-pathing configuration is set to Round Robin Path selection. When you see significant imbalances, you should check to make sure that your ESXi hosts are set to use Round Robin path selection.
Avoiding System Bottlenecks and Imbalances in Your Dell EMC Unity
Even when you follow the best practices you may find that from time to time you have system bottlenecks or imbalances.
IntelliMagic Vision is a cloud-based software and services solution designed to help you minimize the risk of performance and configuration impacts in your environment. IntelliMagic Vision automatically assesses your Dell EMC Unity systems for bottlenecks and imbalances and provide expert services to help you resolve the issues.
This solution is ideal for today’s lean staffing and instantly enhances your storage management’s performance analysis expertise by providing both numeric and stoplight themed risk assessment and teaching text at each step along the analysis path.
Subscribe to our Blogs
Subscribe to our newsletter and receive monthly updates about the latest industry news and high quality content, like webinars, blogs, white papers, and more.
Determining the cause of high I/O latency on your NetApp c-mode system
In this short blog we examine a NetApp c-mode system and perform a root cause analysis to identify the system as having a high CIFS and NFS latency.
5 Things Every Storage Professional Should Be Checking - Prevent Fires Before They Start
This blog identifies five key areas that every storage professional should regularly check to identify and eliminate upcoming issues and, in turn, save time in the long run.
Overcoming SAN Visibility and Management Challenges with Operational Analytics
This white paper walks you through increases in SAN storage complexity along with other modern challenges and how to overcome them using deep SAN analytics.